By Greg Allan, tradesman and creator of Data First Coaching
Why can a stonemasonry business ask AI a question and still get the wrong price?
Most AI tools a stonemason tries first are conversational: type a question, get an answer. The office of a stonemasonry business is scattered across the yard, the van and the workshop, with slab prices from the wholesaler, templating notes from site, CNC cut lists and sealing job sheets held in different places. A chat tool has no access to any of that, so its answer is a guess dressed up as a fact.
Ask it the lead time on a kitchen benchtop install or whether a particular engineered stone is in stock, and it answers with confidence anyway, because a confident wrong answer is still an answer. For a trade where pricing on natural stone and engineered stone moves with the wholesaler, and where a job can involve templating, fabrication, cutting, polishing and install all in one week, that gap between a clever reply and a correct one costs real time to unpick.
Why can updating one slab price mark every quote as wrong?
Picture a stonemason whose yard updates the price of a popular natural stone slab after the wholesaler changes its list. The office changes the number in one place and assumes the job is done. But an AI tool connected to the quoting side was still reading an older price list it believed was current, so every quote using that slab, benchtop fabrication, cladding, paving, carried the wrong figure until someone noticed.
At the same time, measurements and cutout notes taken from a phone during templating on site were never connected to the business's central records, so that work simply disappeared and had to be remeasured back at the workshop. Nothing in the system was wrong on purpose. It was reading an old picture because nothing told it where the current one lived.
What is the difference between AI that chats and AI that works for you?
There are two different things both called AI, and the gap matters more to a stonemason than to most trades. Conversational AI answers whatever you type. AI agents, software that does a job the way you would, with your prices, wording and rules, go and actually do the work: pull the current slab price, draft the job note, keep the quote live as stock and costs change.
For a business running natural stone and engineered stone side by side, with CNC cutting, polishing and install all moving at different speeds, an agent connected to your actual pricing and job records is doing something a chat tool never can. It stays current because it is reading the same numbers your office is working from, not a copy from last month.
Can AI turn your templating notes into a job ready to check?
Yes, and this is one of the clearest uses of AI agents for a stonemason. Rough notes taken on site during templating, stone type, dimensions, cutouts, edge profile, can be turned into a drafted job record automatically, ready for a quick check against what was actually measured, instead of someone back at the workshop retyping a phone note from memory.
The templater still does the templating and still knows the stone. The agent does the rewriting and the matching against the job file, so what was measured on site lands in the business's system the same day, not whenever someone finds time to type up a phone note.
Can AI hold your workshop's how-to in one place?
Yes. A stonemasonry workshop runs on know-how that usually lives in one or two heads: how a sandstone restoration is approached, the sealing process for a particular engineered stone, the silica dust control steps for cutting and polishing, how a memorial is finished. AI agents can turn that know-how into linked help documents that sit in one place, so the answer is there when it is needed, not repeated from memory each time.
That matters most where safety is involved. Dust and silica controls in the workshop are not something a new apprentice should be guessing at from what they half remember someone saying last year. Compliance in this trade is serious and general in its requirements, but the record of how your own workshop handles it should be exact, not passed down by word of mouth.
Should you rent software built for every trade or build on your own data?
Off-the-shelf job management software is built for every trade at once, so a stonemason often ends up working around features meant for a sparky or a landscaper rather than for slab pricing, CNC job sheets or templating. Some of those tools genuinely suit parts of a stonemasonry business, and the business's own data shows which ones earn their place.
Systems built on your own records, shaped around how your jobs actually move from quote to template to cut to install to restoration or sealing follow-up, give AI agents something real to work from instead of a generic template built for nobody in particular. A data first look at where a quote slows down, or where a job note goes missing between site and workshop, shows which ready-made tool is worth keeping and where only a system built around your own business will close the gap.
How does Data First Coaching approach this?
Data First Coaching starts with the stonemasonry business's own numbers before any conversation about AI: quoting speed, margin by job type, benchtop against cladding against restoration, and where job information actually goes missing between site and workshop. Through the Trade Business Intelligence, Greg Allan looks at where the real gap sits, which is often not where the owner first points, and works out whether an AI agent belongs there at all. Greg puts it plainly: “Instead of our ops manager sitting there and just doing the manual task of taking the work notes, rewriting them all.” The how, the build and the setup is worked through inside the Academy, with members, because the Trade Business Intelligence is powered by every member who is open minded, innovative and willing to share what they find.

